Privacy & security at a glance
- Your data is never used to train models.
- Personal data (PII) and sensitive company data is sent only to models running in our own cloud production environment, which are configured for zero data retention. The infrastructure providers are listed in the subprocessor list in our Trust Center.
- Compliant with the EU AI Act, classified as limited risk.
What are you using the data for?
Our goal is to provide you with the best real-time sales recommendations on 1) which companies to target, 2) which prospects to reach out to, and 3) when, how, and what to write.
What data of ours do you use?
This depends on which integrations you connect. For the exact data each one uses, see the integration-specific implementation guides — for example, Salesforce: UserGems Implementation Guide - Salesforce CRM.
Your data is never used for any other purpose, and never for any other customer.
Can our data leak into a trained model, or reach another customer?
No.
- We do not use customer data to train models.
- We use your data only to generate results for you (for example, scoring your target accounts). That processing runs on your own data each time; it is never used for other customers and is never built into a trained model.
- Where personal data (PII) or sensitive company data is processed by AI, it is sent only to models running in our own cloud production environment, which are configured for zero data retention.
How do you generate tailored outbound emails?
Information you provide directly is used in the prompts:
- Your company's value proposition.
- Your buyer personas, with their specific pain points and the value proposition for each.
We use a range of signals to find the right prospects and set the right context for each email. For example, we might identify a prospect who recently changed jobs and instruct the model to congratulate them on the new role.
Signals fall into two categories — prospect signals and account signals — and each can be either a UserGems-native signal (derived from public data we gather) or a first-party signal (derived from your own data, when you connect the integration). See the full list of available signals: https://help.usergems.com/signals.
All of this relevant information about your product, accounts, and prospects is turned into a detailed, individual prompt through a thoroughly fine-tuned and tested process (see data safeguards above).
Compliance and more information
UserGems' use of AI is compliant with the EU AI Act and classified as limited risk.
Our AI produces recommendations — account and prospect scores, rankings, signal suggestions, and draft messages — to support your team's decisions; it does not make automated decisions about individuals on your behalf. A person on your team always reviews the output and decides whether and how to act on it (for example, you review and edit a suggested email before it is sent), and AI-generated content is always presented to you as a suggestion. This human-in-the-loop design aligns with the transparency expectations for limited-risk AI systems under the EU AI Act.
For broader data-protection and security details, see our Trust Center, including:
- Our AI policies and how we enforce EU AI Act compliance
- How we handle personal data and data-subject rights under GDPR / CCPA — including the data of prospects and other individuals
- Our subprocessor list and where data is processed (data residency)
- Data retention and deletion
- Encryption and other technical and organizational security measures
- Our Data Processing Addendum (DPA) and other compliance documentation